For the student of Late Roman and post-Roman Britain the fate of the Roman towns is a recurrent theme. In this paper I propose to examine the final stages of one range of towns, normally now called ‘civitas capitals’, since discussions of the end of Roman towns have normally concentrated on these sites. In what follows ‘town’, unless otherwise modified, refers to these settlements. Even Reece, to judge from the citation of examples in his paper (1980), is principally considering the regional centres, the civitas capitals, and not the so-called ‘small towns’, some of which are indeed larger in area than the smallest of the civitates (for an excellent modern summary of the types and their development see Millett 1990, 65-126 and 145-51). Gildas, too, has these sites in mind (Ruin of Britain 3.2), and controversy is concerned with the eclipse of these once important ‘public towns’, rather than with the much more obscure fate of the small towns, for which one may now consult Burnham and Wacher (1990). I shall be arguing that our interpretations of the Late Roman civitates misrepresent their nature when they are discussed in economic terms; that to consider them as social units reconciles the controversy between Reece’s position and more traditionalist notions of the town and allows us a clearer picture of the transition between the Roman province and the English successor kingdoms.